Project I: This project will provide a rare opportunity to exploit several large databases that contain a wide range of drug use and psychosocial data on four rural, ethnic populations. American Indian, Mexican, American, Black American and White American adolescents attending school. Additional data will be available to compare and contrast these populations with non-rural youth of ethnic background. All of the studies from which the data are taken use the same instrumentation and survey methodology and, with one exception, use stratified random sampling. The intent of this project is to identify those factors that have strong implications for prevention and to make recommendations about prevention strategies. The data will come from projects in progress within the Tri-Ethnic Center for Prevention Research. Data collection will be ongoing, providing the opportunity for an array of analyses to be conducted for cross validation of all findings. In addition to cross validation, each analysis will take place by gender, grade in school and ethnicity to produce an extremely rich description of drug use and interactions among these variables and populations. Several analytic approaches are proposed. First drug use prevalence rates and patterns of use across populations will be compared to determine whether personal and demographic characteristics lead to differences. Second, dependent and independent variables will be examined across ethnic groups to establish reliability and validity estimates. EQS using confirmatory factor analysis will be employed to test for measurement invariance. Third, mean differences on psychosocial variables will be compared and a series of EQS models will be generated and differences elucidated A central element of these models will be to detect the impact of cultural identification on substance use. Finally, throughout the course of the project findings will be routinely published and made available to both professionals and practitioners.